Fault Tolerant Model for Redundant System Based on Health Assessment

Abstract

Fault tolerance with redundant module is one of the effective ways to improve the reliability of complex system. Most of the traditional redundant algorithms require all redundant components to work at the same time to vote. It requires a lot of resources, and it is impossible to know the specific health status of the working module. In order to solve this problem, we propose a redundant fault tolerant system based on neural network health assessment. In this model, one-dimensional convolutional autoencoder is trained to assess the health state of the module. According to the health score, it can judge whether the working module is faulty and whether it needs to switch to the redundant module. The model is validated in the bearing fault data set of Case Western Reserve University. The experimental results show that the model can effectively identify the health status of the module, and the fault tolerance of redundant system can be effectively realized through module switching.